But the results of the initial human trial also indicated two doses of vaccine might be necessary to provide long-lasting immunity.

If true, that will make bringing the pandemic to an end both more logistically challenging and more expensive, since a large percentage of the world’s population will need to be injected two different times.

By partnering with researchers at the University of Oxford, which began developing a COVID-19 vaccine as soon as the virus’s genetic code was published in January, AstraZeneca leaped ahead of competing drugmakers, many with better track records in vaccines, to gain the inside track on delivering a vaccine that could help end the global pandemic.

AstraZeneca has made arrangements to provide about 2 billion doses of Oxford’s vaccine, including 300 million doses for the U.S. and 100 million for the U.K. It has agreed to provide 400 million doses to the European Union and made arrangements with partners to provide 1.3 billion doses for low- and middle-income nations.

But even so, it will not be enough, the company’s chief executive officer, Pascal Soriot, said in a media call on Monday.

“We may need two doses,” he said. “Other vaccines will need two doses. So we will need several vaccines.”

AstraZeneca is providing these doses at cost. It will only make a profit if, once the pandemic has ended, COVID-19 becomes a recurrent threat for which people need regular booster vaccinations—similar to the seasonal flu—and if its vaccine proves the most effective at combating it.

There are currently at least 163 different COVID-19 vaccine candidates in some stage of development, according to the World Health Organization. But the Oxford vaccine is arguably the furthest along in large-scale human clinical trials.

The research published in The Lancet comes from part of the first cohort of 1,077 people vaccinated in the U.K., as part of Oxford’s Phase I safety trial.

Larger second- and third-stage clinical trials—which are designed to prove how effective the vaccine may be—are currently underway in South Africa, Brazil, and the U.K., involving some 17,000 people. AstraZeneca is planning to begin a stage-three clinical trial in the U.S. involving 30,000 people next month.

AstraZeneca and the Oxford researchers had initially hoped that the results of these late-stage clinical trials might be available by the end of August. But falling infection rates in many parts of the world may mean it will take longer to prove that the vaccine is effective. Soriot said that results should be clear sometime between September and November and that the company is ready to deliver the first doses as soon as regulators approve the vaccine, which they are expected to do on an emergency basis if the late-stage clinical trials are successful.

Vaccines have traditionally been made using proteins that come directly from the virus they are meant to protect against. Either a weakened version of that virus or fragments of the virus are used. The Oxford vaccine is different: It uses a harmless chimpanzee virus and genetically modifies it so that it produces the same spike protein found on SARS-CoV-2, the coronavirus that causes COVID-19.

The clinical trial showed that Oxford’s vaccine was safe, with most of those inoculated experiencing only mild symptoms from the inoculation, similar to those patients sometimes experience with a flu shot, such as fatigue, a headache, malaise, and a mild fever. None of those given the vaccine experienced a severe reaction.

Although the study was primarily designed to prove the vaccine’s safety, not its effectiveness, the researchers published their study of the immune response in 25 of those who received the single vaccine dose, and 10 others who received two doses.

Mene Pangalos, AstraZeneca’s executive vice president of biopharmaceuticals research and development, said that this immune response data was limited by the time it takes to run the assays necessary to analyze the blood samples. He said the Oxford researchers planned to publish additional blood data from more of the Phase I volunteers as it became available.

In the published results, of those who received a single dose, 91% developed neutralizing antibodies. In those who received two doses, 100% developed these antibodies. The antibody levels peaked at 28 days after injection and remained elevated throughout the 56 days of the study period.

The researchers found that those who received two doses of the vaccine had antibody levels similar to those found in people who become ill with COVID-19 but recover.

In addition, the vaccine triggered another immune response involving T cells. These killer cells seek out and attack pathogens in the body, and some of them can remember viruses they’ve battled before, helping the body to quickly dispatch them. All of the patients studied showed a strong T-cell response, peaking 14 days after inoculation, even from a single dose of the vaccine.

It is not known how long antibodies to COVID-19 remain active. But Pangalos said that based on Oxford’s work with a similar vaccine for MERS, which conferred immunity for 18 to 22 months after two doses, he is confident that, if it is proven effective, this COVID-19 vaccine would also, with two doses, confer immunity for at least a year.

He also noted that the Phase I trial showed that the mild symptoms people experienced in reaction to the vaccine actually lessened with the second dose, which meant people would be able to tolerate getting multiple booster shots of the vaccine throughout their lives.

More coronavirus coverage from Fortune:

• Flights with empty middle seats decrease COVID risk 79%, study says
• Welcome to the “cat bond” market, where pandemics and hurricanes can mean big returns—or total wipeout
• How chatbots are helping in the fight against COVID-19
• From ball gowns to medical gowns: How one designer changed her business model to help local health care providers
• What the post-pandemic housing market might look like, according to the CEO of Century 21